In September 2020, China predicted that its CO₂ emissions will peak by 2030 (China Government Network 2021). Energy storage technologies are yet to develop completely, and a major proportion of the energy requirements is still met using fossil fuels. Accordingly, it is urgent to realize the efficient and clean utilization of coal (Yang 2022; Chai 2021). Coal gasification technology is one of the main methods in this regard (Wang 2021a, b; Cao 2021; Ding et al. 2022; Smolinski et al. 2022; Zhakupov et al. 2022). However, gasification generates a large amount of solid waste, known as fine slag (FS), which is usually disposed in landfills and causes environmental pollution. This limits the development of energy-efficient coal gasification technologies. Nonetheless, the inadequate utilization of the high residual carbon content in this FS is a waste of resources (Liu 2021).

Figure 1 presents the process for producing FS in a Shell gasifier. Raw coal particles enter the gasifier and undergo rapid pyrolysis. In this process, the raw coal particles undergo expansion and rupture, carbon matrix particles are consumed, and tiny ash particles are formed. Some of the particles on the gasifier wall collected from the molten slag stream downwards along the furnace wall and settle into the slag removal system. In the slag removal system, molten slag disintegrates into solid particles of various sizes. In industries, this portion of solid particles is known as coarse slag. The other portion of the particulates moves from the top of the gasifier to the raw synthetic gas. The fly ash generated from washing or trapped by dust removal devices is captured by the ceramic filter. The untrapped parts passes to mix into the black water in the cooling chamber and is transferred to the filter press to form the FS after an empty flash (Wang 2021a, b; Shen 2021; Zhao 2010).

Schematic of Shell coal gasification

Research on FS is mainly focused on its physical and chemical properties. The FS mainly contains SiO₂, Al₂O₃, CaO, and Fe₂O₃ (Lv 2021). The inorganic components include crystalline components (silicate, aluminosilicate, calcium, iron, and iron oxide) and glass components (calcium aluminosilicate glass) (Wu 2015). FS particles with different sizes have different characteristics, mainly because they undergo different processes in the gasifier. The breakdown of particulates, agglomeration of ash, and slag deposition should be considered (Pan 2016). Xu (2009) found that the inorganic substances were primarily spherical, while the residual carbon existed in a free flocculated form. Miao (2020a, b, c) demonstrated that the rich ore particles in the FS comprised abundant particles with small pores and a small number of solid particles with large pores. The large pores were independent of each other within the particles, and there were cracks, holes, and circular holes on the surface of the particles.

To further reveal the structural characteristics of the FS, many researchers have focused on the chemical structures of the residual carbon and inorganic components in the FS. In addition to the discrete residual carbon particles in the FS, there are two other types of residual carbon particles. In the first type, the residual carbon particles form a matrix of rich ore particulates, while in the second type, the residual carbon is chemically and physically bound to inorganic matter in a rich granular matrix. The most important chemical bonds of the residual carbon in the rich granular matrix are C–C, C–H, and C–O. C–O combines with inorganic elements in the rich mineral particulate matter to form C–O–M groups (M represents inorganic elements) (Miao 2020a, b, c). Thus, there might be a specific relationship between the amounts of inorganic constituents and residual carbon in the FS. Melting test of the fine carbon slag mixture confirmed that the inorganic components tended to agglomerate into larger carbon-free spheres when melted in the carbon matrix (Miao 2020a, b, c). The smelting phase formed by the particles of the ore-rich materials in the gas generator covered the carbon materials in the particles (Krishnamoorthy 2017). The fate of coal particles can be analyzed in terms of coal migration, drywall interaction, coal deposition, and heterogeneous gasification response (Troiano 2018).

Thus, coal gasification is a complex process, and many factors influence the formation of the FS. The pressure and temperature fluctuations in an actual operation of the gasifier affect the morphology of the FS. The mechanism of interaction between the residual carbon and inorganic components of the FS is still elusive, and there is no suitable method to explain the formation of FS in the furnace. In this regard, it is necessary to analyze the macroscopic and microscopic morphological characteristics of the FS, which will be beneficial for the resource utilization of FS and the operation efficiency of gasification plants. This was one of the prime objectives of our study—we collected FS from actual factories, and the features of the FS were analyzed at the macroscopic and microscopic scales. The macrostructure and microstructure highlight the multiscale morphological characteristics of the FS. We also speculated the possible mechanism of FS formation based on the characteristics of FS. Additionally, the application prospects and contribution of the analysis of these multiscale factors of the FS are presented.

2.1 Samples preparation

Different FS specimens were collected at a sampling interval of 2 h and evenly mixed to obtain the test samples. Table 1 shows the proximate and ultimate analysis of FS. Five specimens of FS (FS1#, FS2#, FS3#, FS4#, and FS5#) are obtained from a Shell gasifier in China. The gasification unit was in regular operation. The immediate and final analyses of the FS samples revealed the presence of numerous inorganic and carbon components. After comparing the properties of different FS, we selected FS1#, FS2#, and FS3# for further characterization at different scales. FS was burned to a constant weight at 815 °C to produce the ash of fine slag (FSA).

Various components of FSA are present as oxides. The highest content is that of SiO₂, followed by Al₂O₃, CaO, and Fe₂O₃. MgO, Na₂O, K₂O, and P₂O₅ are present in small amounts. The oxide contents are shown in Fig. 2. Simultaneously, we analyzed the chemical composition of the coal ash in the raw coal (RC) formed during the operation of the gasifier. There is a correspondence between RC and FS. Analysis of the chemical composition of the raw coal ash (RCA) indicates enrichment of the iron content in the FS.

Chemical composition of FSA and RCA

2.2 Experiment

The FS is pretreated with the aim of sufficiently dispersing the particles, and the pre-processing was divided into three steps:

Step 1: Addition of 100 g of FS in 200 g of anhydrous ethanol.

Step 2: Ultrasonic treatment for 2 h.

Step 3: Drying at 40 °C for 8 h in a nitrogen atmosphere.

The physical and chemical properties of the pre-processed FS were characterized at the macroscopic and microscopic scales. The experimental flowchart is shown in Fig. 3.

Experimental flow chart

2.2.1 Macroscopic characterization

For macroscopic characterization of the FS, the particle size distributions and apparent morphologies were examined. The particle size was analyzed using a laser particle size analyzer. FS (50 mg) was completely dispersed in alcohol (dispersant) and placed into the instrument for analysis. The particle size distribution was determined from the average of three experiments. The repeatability error and accuracy error were < 1%.

The apparent morphology of the FS was analyzed by Scanning Electron Mictoscopy (SEM, FlexSEM1000, Hitachi, Japan). The sample was evenly distributed over a conductive adhesive, covered with a conductive film, and placed in the sample chamber. The sample chamber was vacuumed, and a high voltage was applied during the experiment. Typical characteristics were analyzed at various magnifications. ImageJ was used to analyze the images and determine the dimensions of the species.

2.2.2 Microscopic characterization

The crystal mineral compositions of the FS and FSA were analyzed by X-ray diffraction (XRD, Ultima IV, Neo-Confucianism Corporation, Japan). Samples were scanned in the 2θ range of 15°–80° with a step size of 0.02° and scanning speed of 2°/min using Cu Kα radiation (λ = 0.15406 nm).

The carbon structure was analyzed by Raman spectroscopy (Raman, Renishaw-Invia, Renishaw, Britain). After further grinding, the samples were placed into the sample table. A 532-nm laser was used for excitation to scan the samples in the range of 800–2000 cm⁻¹.

The functional groups in the FS were analyzed by Fourier transform infrared (FTIR, Nicolet iS50, Thermo Fisher Scientific, The United States) spectroscopy. Spectra were obtained in the range of 400–2000 cm⁻¹, and most bands were observed in the mid-infrared region of the spectra, corresponding to inorganic functional groups. The samples were prepared in the form of KBr pellets, with a sample to KBr mass ratio of 1:100.

The surface elements were investigated by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha, The United States), using Al Kα radiation (HV = 1486.6 eV) as the excitation source. Full-spectrum scanning was performed at a pass energy of 100 eV and step size of 1 eV. Narrow-spectrum scanning was performed at a pass energy of 50 eV and step size of 0.1 eV.

3.1 Geometrical dimensions of FS at a macroscopic scale

The particle size of the pretreated FS was analyzed, and the results are shown in Fig. 4 and Table 2. All the three types of FS exhibited the characteristic “three-peak distribution,” although it was the most prominent for FS3#. This indicated a certain coherence for FS formation. The overall particulate size ranged from 0.5 to 350 μm. The first, second, and third peaks ranged from 0.5 to 5 μm, 5 to 40 μm, and 40 to 350 μm, respectively, with the particle sizes of the samples overlapping.

Particle size distribution curves

In Figs. 5, 6, and 7, the local areas were randomly chosen to determine the global particle statistics. The particle sizes varied and were scattered throughout this region. The macrostructure of the FS produced was divided into two main categories. In the first category, single spherical particles of various sizes were observed—the largest spherical particle diameters in Figs. 5, 6, and 7 were approximately 45, 45, and 63 μm, respectively.

Overall particle size distribution in FS1#

Overall particle size distribution in FS2#

Overall particle size distribution in FS3#

Figure 8 shows SEM images at various magnifications. A-1 and B-1 show bulk particles with sizes ranging from 200 to 300 μm. In A-2, A-3, B-2, and B-3, the surface contained cracks or holes of various sizes, and the surface crack was more significant. Many small irregular and spherical particles were attached onto the particle surface through both physical adsorption and engraving.

Apparent particle morphology at different magnifications

C-1 and D-1 show dense particles. Images C-2, C-3, D-2, and D-3 obtained at higher magnifications show that the surfaces contained cracks or folds and were relatively smooth. Compared with A-1 and B-1, the amount of particulate matter attached to the surface was significantly less.

The particulates shown in E-1 are porous and irregular. The corresponding higher-magnification images (E-2 and E-3) show many pores of various sizes on the surface. While the particles in E-1 had the same porosity characteristics as those in A-1 and B-1, there were no small charged particles on their surfaces. Small particles were physically attached to such particles, and the attached particles were dispersed by vibration during the pretreatment process.

Images F-2, F-3, G-2, and G-3 show visible fold patterns on the surfaces. The overall particles in F-3 appear folded, whereas the localized particles in G-3 appear folded, with some small irregular as well as spherical particles embedded in the fold. Crinkled particles can be formed due to cracks in high-temperature coal particles. Such particles have small local cracks that expand into larger cracks. In conclusion, the number and size of the spherical particles attached to the surface and the pore structure of irregular particles are significantly different because of the different structures of the irregular particle matrix. The formation process of FS is related to particle breakage, ash accumulation, and slag deposition, and hence, FS with different particle sizes may have different characteristics.

Figure 9 shows the morphology of the fine particles, whose size was less than 40 μm. The smaller particles (< 5 μm) were attached to their periphery or surface.

Apparent morphology of particles with sizes less than 40 μm

In Fig. 9, the irregular particles in the images are either dense particles (I-1 and I-5), or cracked particles (I-4 and I-6), or a combination of acicular particles, floc particles, and glitter particle (I-2 and I-3). The primary sources were tiny particles of unreacted or partially reacted raw coal and the reactions during the gasification process.

Images S-1 to S-6 show spherical particles of different diameters. Large spherical particles were not smooth, and some spherical particles had a rough surface (S-1, S-2, and S-4). Certain smooth surfaces were scattered with patches of various sizes (S-3). The spherical particles also had partial depressions (S-3) or the precursors (S-5) on the surface, indicating that the spherical particles were not formed directly but were formed through specific physical and chemical reactions.

3.2 Characteristics of the carbon-containing components at the microscopic scale

Figure 10 shows the XRD patterns of the FS. There was no obvious signal of mineral diffraction from the FS. Peak 002 (approximately 25°) and peak 100 (approximately 45°) in the diffraction pattern indicate a graphite-like structure in the FS. During gasification, the working temperature was higher than the melting temperature of the coal ash. Consequently, the mineral melted into a glassy material consisting mainly of fused silicate and aluminum silicate. After fast cooling, the slag did not exhibit a high degree of crystallinity and formed more amorphous components.

XRD patterns of the FS

Raman spectroscopy is an effective technique for characterizing the structure of carbon owing to the sensitivity of Raman signals to crystalline and amorphous structures. Figures 11a, b, and c shows characteristic bands at approximately 1350 cm⁻¹ (peak D) and 1590 cm⁻¹ (peak G) (Yu 2021). Deconvolution of the spectra showed five peaks (G, D1, D2, D3, and D4). Peak G corresponds to an orderly network of graphitic carbon. Peak D1 (1350 cm⁻¹) corresponds to a more extensive aromatic nucleus system related to edges or other defects (such as edge carbon atoms or heteroatoms) and originates from the vibrations in the isolated sp² hybrid bond planes. Peak D2 (1620 cm⁻¹) appears simultaneously with peak D1, reflecting the vibration mode of the graphite lattice with E_2g symmetry. Peak D3 (1520 cm⁻¹) can be attributed to a minor aromatic nucleus system in the aromatic structural unit, corresponding to the sp² hybrid structure of amorphous carbon in the organic molecule or functional groups. Peak D4 (1200 cm⁻¹) is usually observed for combined polyene structure with sp²–sp³ hybrid systems owing to crystallite edges or combined C=C and C–C bonds, typically representing the active site, reactive site, and cross-linking structure (Xu 2021; Wu 2014).

Raman spectra and composition of FS

The A_D3+D4/A_All ratio represents the number of reaction sites on the carbon structure. The A_D1/A_G and A_G/A_All ratios represent the magnitude of the carbon structure, and the A_D3/A_G+D2+D3 ratio represents the amount of amorphous carbon structure. The proportions of these structures are shown in Fig. 11d. According to the fitted peak areas, the areas decrease in the order A_D1 > A_D3 > A_D4 > A_G > A_D2. The highest area of peak D1 implies a high number of aromatic nuclei, mainly from the graphite structure, with a more intense aromatic structure. This peak D3 originates primarily from the hybrid sp² structure of amorphous carbon. Peak D4 reflects the presence of some active sites, indicating that the carbon structure still has specific reactivity. The active sites were derived from the original coal particles that did not respond completely to the gasification process. According to the A_D3+D4/A_All and A_D3/A_G+D2+D3 ratios, the number of reaction sites in FS1#, caused by the carbon structure, was higher. Fundamental analysis revealed that FS1# has more volatiles. The A_D1/A_G and A_G/A_All ratios indicate that FS2# and FS3# have a higher-order structure compared to FS1#. The more the number of reactive sites on amorphous carbon, the limited will be the consumption or transformation in carbon gasification, and higher will be the amount of carbon structures preserved in the FS (Guo 2022). It is evident that the structure of the residual carbon in FS has specific commonalities. This is directly related to the nature of the raw coal and the course of the reaction owing to the difference in the degree of crystallinity.

XPS is used to determine the elemental composition, chemical state, and electronic state of materials (Perry and Grint 1983). The C 1s XPS spectrum ranges from 281 to 292 eV. The carbon on the FS surface consists of five components. The C–C and C–H peaks appeared at 284.8 eV, while hydroxyl/ether (C–O), carbonyl (C=O), carboxyl (COO⁻), and π–π* peaks appeared at 285.4 ± 0.2 eV, 287.5 ± 0.2 eV, 289.0 ± 0.2 eV, and 291.5 ± 0.2 eV 3233, respectively. The peak at 294.3 eV can be attributed to the oscillations involving π–π* excitation in unsaturated and aromatic species (Pantea 2003; Kozowski 2004). The peak deconvolution is shown in Fig. 12.

Fitted peaks and parameters of the C 1s spectrum of FS

The peaks were fitted to obtain the area under the peak, and the ratios of the different peak areas were obtained (Fig. 12d). The proportion of aromatic C–C or C–H bonds on the surface of the FS was the highest. Carbon primarily formed a single bond (hydroxyl or ether), although small amounts of double bonded carbon (as C=O and COO⁻ groups) were also present. In raw coal, most C=O and COO⁻ bonds dissociate after the various high-temperature processes. At elevated temperatures, the C–O groups in phenol, alcohol, or other compounds formed are easily dissociated. This is probably because the residual C–O groups did not react entirely with the raw coal particles; a portion of these groups is present in the carbonate structure of the inorganic components (Miao 2020a, b, c; Li 2021; Xie 2016).

On the one hand, coal particles undergo complex reactions in the gasifier, in which the carbon structure will change under the influence of temperature. On the other hand, the residual carbon in the FS may be formed when the coal particles in the gasifier do not reach the required conditions for a complete reaction or need more reaction time. Carbon-containing components' bond binding and occurrence forms show relatively consistent characteristics. Although there are minor differences, the degree of common traits is far greater than the differences. According to Raman results, the area of peaks D3 and D4 in F1# is more significant, which reflects the higher content of the hybrid sp² structure of amorphous carbon and crystal edge or the combination of C=C and C–C bonds. In XPS results, there are relatively more C–C bonds in F1#. These results indicate that FS1# has more reaction sites and cross-linking structures than FS2# and FS3#.

3.3 Characteristics of the inorganic components at a microscopic scale

To understand the primary crystalline forms of FSA, the mineral composition of the crystals of various FSA was analyzed by XRD. The XRD patterns are shown in Fig. 13. The spectra of the three FSA are basically the same, indicating common features of the slag after combustion. The prominent peaks correspond to hematite (Fe₂O₃), anhydrous gypsum (CaSO₄), and mullite (Al₆Si₂O₁₃). Clusters of peaks were detected at a 2θ value of 25°. The FSA contained a certain amount of amorphous inorganics. Combining with the XPS results, it was concluded that the amorphous structures contained silicon, aluminum, iron, and calcium (Zhang 2019).

Mineral composition of the FSA

The FTIR spectrum was not affected by the crystalline structure of the mineral constituents. However, there were numerous amorphous mineral components in the FS obtained under the high-temperature atmosphere of the gasifier. A portion of the mid-infrared region (400–2000 cm⁻¹) was selected for the experiment. Figure 14 shows the spectra of the three types of FS. The spectra of the three FS were similar. Thus, all the FS have similar functional groups, although there are minor differences.

FTIR spectra of FS

Bands at 1585 and 1641 cm⁻¹ can be attributed to aromatic C=C stretching vibrations of graphite carbon. The FS has a particular aromatic structure about the residual carbon (Guo 2020). The bands from 1138 to 1009 cm⁻¹ are asymmetrical, corresponding to Si–O–Si. The spectra showed a weak shoulder at 710–770 cm⁻¹, attributable to the stretching vibration of Si–O–Al (720 cm⁻¹) and symmetrical stretching vibration of Si–O–Si (770 cm⁻¹). The signal from the Si–O flexural vibration in quartz appeared at 470 cm⁻¹ in the spectra (Zhao 2021; Mozgawa 2014). The bands from 1138 to 1009 cm⁻¹ are asymmetrical, corresponding to Si–O–Si. The spectra showed a weak shoulder at 710–770 cm⁻¹, attributable to the stretching vibration of Si–O–Al (720 cm⁻¹) and symmetrical stretching vibration of Si–O–Si (770 cm⁻¹). The signal from the Si–O flexural vibration in quartz appeared at 470 cm⁻¹ in the spectra (Lin 2012). The peak at 612 cm⁻¹ mainly originated from the in-plane bending vibration of SO₄²⁻ and could be attributed to the molecular vibration of anhydrite (Yin 2018). The FTIR spectra of FS showed an obvious peak for the Fe–O bond and a specific vibrational stretching pattern near 452 and 565 cm⁻¹ (Chen 2017; Yadav 2020).

As shown in Fig. 2, there was an obvious enrichment of elemental Fe. XPS was used to further analyze the oxidation state of iron in Fig. 15. The Fe 2p_1/2 and Fe 2p_3/2 peak positions depend on the ionic state of Fe. The satellite peaks of Fe 2p_1/2 and Fe 2p_3/2 are also dependent on the oxidization state and have been used to qualitatively determine the oxidation state of iron ions.

In the high-resolution Fe 2p spectrum, peaks appeared at 724.4 ± 0.2 eV (Fe 2p_1/2) and 710.8 ± 0.2 eV (Fe 2p_3/2). The satellite signal appeared at 719.2 ± 0.2 eV and typically indicates the transformation of FeO (Fe²⁺) to Fe₃O₄ and Fe₂O₃ (Fe³⁺) to Fe₃O₄. Meanwhile, the Fe 2p_3/2 peak can be deconvoluted into Fe(III) peak at 711.0 ± 0.2 eV and Fe(II) peak at 713 ± 0.2 eV (Yang 2015; Ai 2013). The specific peak fitting results are shown in Fig. 15A, B, and C. Iron compounds can be either high-spin or low-spin (as defined by the crystal field theory). The spin of Fe(III) compounds always increases, leading to complex multiple-peak Fe 2p spectra. Fe(II) compounds may have high or low spin (Biesinger 2011).

Peak fitting of the Fe 2p XPS spectrum and composition of Fe in the FS

Figure 15d shows the percentage areas of Fe(II) and Fe(III). Fe mainly exists in the form of Fe(III). The crystalline mineral composition of FSA reveals the presence of hematite with trivalent iron. Simultaneously, oxides containing divalent iron or other inorganic minerals can be fused with elements such as silicon and aluminum. The form of iron in the FS can change during the gasification reaction, although there are some common features. The state of Fe in the FS still requires investigation.

It can be seen that both crystal mineral composition and chemical bond display relatively consistent characteristics. While there are minor differences, the degree of common traits is far greater than the differences. There are two main reasons for the difference: (1) The production of FS is directly related to the actual operation condition and is subject to the temperature, pressure and the properties of the raw coal in the furnace: (2) The FS mainly contains inorganic and carbon-containing components, among which there is a specific occurrence relationship between carbon and ash. The difference may be due to the difference in carbon-containing components.

Figure 16 summarizes the partial characteristics of the FS at different scales. The FS primarily contains large amounts of inorganic and carbon components. Macroscopically, FS exhibits a characteristics “three-peak distribution.” FS particles exhibit regular spheroidal particles and irregular particles. Microscopically, the carbon structure has a certain degree of order. C 1s contains many aromatic structures. FS contains a vitreous inorganic component. The inorganic components are primarily silicon, aluminum, iron, and calcium. Inorganic components mainly contain Si–O–Si, Si–O–Al, Si–O, SO₄²⁻, and Fe–O bonds.

Analysis of partial characteristics of FS at different scales

In conclusion, certain properties, such as particle size distribution, composition, crystalline structure, and valence binding, were common in the FS. The reaction of coal in the gasifier is complex, and the FS in the Shell gasifier has specific differences, although many common characteristics can be observed. The FS primarily contains large amounts of inorganic and carbon components. The inorganic components were mainly silicon, aluminum, iron, and calcium. Macroscopically, FS exhibits a characteristics "three-peak distribution." Microscopically, the carbon structure has a certain degree of order. C 1s contains many aromatic structures, and A_C–C&C–H > A_C–O > A_π–π* > A_C=O > A_COO⁻. FS contains a vitreous inorganic component. The inorganic components are primarily silicon, aluminum, iron, and calcium, with a total content of more than 90%. Inorganic components mainly contain Si–O–Si, Si–O–Al, Si–O, SO₄²⁻, and Fe–O bonds. Particularly, Fe mainly exist in the form of Fe(III), and the Fe 2p spectrum can be deconvoluted into Fe 2p_1/2, Fe 2p_3/2, and satellite peaks. To develop efficient coal gasification technologies, the training and removal of FS would be a critical research direction in the future. The findings of this study will be beneficial in many aspects. The utilization of FS resources will differ based on the composition and properties. The qualitative use of FS can be better realized after the characteristics of the different components are identified. The mineral composition, inorganic functional groups on the surface of FS, and the structural characteristics of C and Fe revealed in this study will be beneficial for elucidating the formation process of FS.